These are my answers to the general field exam, one of the competency exams for Ph. D candidacy in philosophy of science in Committee on the History and Philosophy of Science (CHPS). The exam is a four-hour written exam, and I had to pick up four questions out of 12 questions. I have corrected grammatical mistakes in the original answers, but I have deliberately left my factual misstatements unchanged, because I think some of you are interested in knowing what kind of answers are considered as passable (Yes, I have passed the exam!).

1. Question: What is or should be the relation between history of science and philosophy of science? Critically evaluate the view on this issue of one of the following: Lakatos, Kuhn, Feyerabend, Losee, Kitcher. Argue for a positon of your own.


What should be the relation between history of science and philosophy of science? I analyze Imre Lakatos's position on this issue.

Lakatos's position in philosophy of science is called "research programme". According to him, scientific changes occurs as choices between two or more conflicting research programmes. A research programme consists of a series of theories, which share a "hard core", i.e. basic claim of the programme. Scientists try to protect hard core from unfavorable evidence by changing auxiliary hypotheses, called "protective belt". This change in auxiliary hypotheses should be "progressive", namely later theories in the programme should make novel predictions which former theories in the programme could not make. Now, when one of the competing research programmes is progressive and all others are degenerative (i.e. , making only ad hoc changes which does not lead to novel prediction), scientists choose the progressive research programme.

Now, what is the status of this claim? Is this normative, in the sense scientists should obey this rule regardless of the actual course of history of science? Or is this descriptive, in the sense it just describe the way scientists chose scientific theories in the history? According to Lakatos, niether of them. This is a rational reconstraction of the history. Namely, Lakatos's view is informed by history of science, but he reconstruct the history so that we can obtain methodological advice from the history.

Therefore, in Lakatos's view history of science and philosophy of science are closely related with each other. When historians write history of science, they should have possible philosophical interpretations of the history in mind, and evaluate and reconstruct historical materials according to the philosophical perspective. When philosophers write philosophy of science, their normative claims should have some basis in history, even though philosophers do not have to find a scientist who have exactly obeyed the normative claim and succeeded.

There are several points to be analyzed in Lakatos's view. First of all, he takes normative nature of philosophy of science pretty much for granted. There are other philosophers, like Losee, who advocate the view that philosophy of science should be descriptive. If Losee is right, we need to expect an even closer relationship between history of science and philosophy of science. But for me, normative nature of philosophy of science seems to be an indispensable part of it. I think that the basic thrust of doing philosophy of science is to find out why science has attained its great success, and to learn from the success useful advice applicable to future investigations and investigations in other fields. If we do not have this motivation, what is the point of doing philosophy of science instead of history or sociology of science? Thus I am sympathetic with Lakatos's view that philosophy of science need to be normative.

However, I have grave doubt if the normative task of philosophy of science is attained by his rational reconstruction method. If we really want to learn from what was going on in the history, we need to be careful. We can easily distort the history from our own prejudice, and miss important aspects of the history. For example, as my summary of Lakatos's view shows, he places a very strong emphasis on novel predictions. When we have this view in mind, and when a theory seems to have made a successful novel prediction, we can be easily lead to the conclusion that the theory was accepted because of the novel prediction. But the history is not that simple. For example, there is an interesting study by Stephen Brush on the reason of scientists' accepting the general theory of relativity. As is well-known, the general theory of relativity predicted the deflection of starlight by the sun, and the prediction was confirmed. But according to Brush, this was not the primary reason why scientists accepted the general theory of relativity. Rather, they were most impressed by the fact that the theory successfully explain the perihelion advance of Mercury, which is not a novel fact at all.

This example of Brush's study also provides a good example of use of philosophy in history of science. In this case, Lakatos's philosophical claim provided a research agenda for the historical research. For this kind of use of philosophy, historians need not to think about if the philosophical view is a rational reconstruction of the history. Historians can take any bold philosophical claim to test in actual historical cases.

As a conclusion, I agree with Lakatos that philosophy of science as a normative task and history of science can be benefited from each other. What I cannot agree with him is his view on rational reconstruction. Philosophy of science need not be a rational reconstruction, and historians can be benefited by a philosophical view which is not a rational reconstruction.

3. Question: Is there a rational way to distinguish "science from other activities such as "pseudo-scinece," politics, literary criticism, religion, ethics, etc.? Discuss some of the more plausible of such proposed "demarcation" criteria, and either (a) defend one that you consider the best, or (b) show that no such criterion is justifiable or possible. In either case, identify and discuss two or three examples of fields whose "scientific" credentials are in dispute.


Demarcation of science from non-science has been a major motivation for some influential philosophers of science, especially for Karl Popper. In this answer, I would like to take two examples of pseudo-science dispute, Freudian psychoanalysis and creationism, to evaluate some of demarcation criteria.

Freudian psychoanalysis is a theory on human psychological structure. To put crudely, According to Freudian psychoanalysis, our mental structure is divided into three parts -- id, ego, and sperego. Superego represses our unconscious desires in our id, and mental deseases are caused by the repression. It also claims that mental diseases can be cured by becoming conscious about the repression. Freudians also claims that the success of the psychoanalytic therapy is a confirmation of Freudian theory. There are objections to these claims. First of all, the theory does not tell us in each case what is the repression which causes the disease of a particular patient. There is a wide room for interpretation, and if a psychoanalyst is competent enough, she/he can construct a story about repression for any case. According to Popper, this is a grave problem for a scientific field, because this means that Freudian psychoanalysis is in principle unfalsifiable. Popper claimed that because of this Freudian psychoanalysis is not science. Freudians' claim about the interpretation of the success of their therapy is also attacked, because other school of psychoanalysis with radically different theoretical framework also succeed in their therapy.

Creationism is another example of pseudo-science debate. The primary claim of creationism is that species never evolve into other species, and each species is created independently. This was a orthodox biological view in the first half of the 19th century, but the debate we concern here is between creationists of present time and biologists (including a philosopher of biology, Michael Ruse). These creationists are fundamentalists in their religious view, and obviously their claim comes from literal interpretation of the Bible. In the trial on the equal treatment of `creation science' and `evolution science' (these are the name used in Arkansas law for the equal treatment in 1980), Michael Ruse argued that creationism is not science now, because creationists ignore all evidence which is not favorable to their claims. Ruse also uses other criteria including Popper's falsifiability, testabilty, tentativeness and fruitfulness. But these criteria are used to support his major argument, the negligence of evidence.

These two cases show that the demarcation is not easy at all. As for Freudian psychoanalysis, as Grunbaum points out, it does make falsifiable claims about patients -- for example, it predicts that such and such patients are likely to have had such and such experience in the past, or that such and such patients will be cured by such and such method, and so on. The problem is, when the prediction is falsified, Freudians just revise their theory to meet new evidence. But here is a problem of Popperian falsificationism itself. The revision of theory by new evidence is what scientists always do. If this is not admissible, then Popperian falsificationism rules out all established scientific fields from science.

Problem with Ruse's argument is rather different. A close examination of his argument shows that the distinction he makes is not between science and pseudo-science, but between scientists and pseudo-scientists. Creationism, as an abstract theoretical claim, can be a legitimate science, if only we have evidence for it. The problem with creationists is that they ignore evidence because of their religious conviction.

So, I think that the attempts to demarcate science from pseudo-science failed in both cases. But the way they failed, especially the way Ruse failed is illuminate. Any claim can be a scientific theory, if suitable evidence supports the claim. So if we look at structure of scientific theories, we never find a demarcation criteria. Rather, what we really need to worry about is the way practitioners in the field treat the theory and evidence. If Freudian psychoanalysts change their theory too light-mindedly, or if they ignore evidence, this is a problem for a scientific field. And I think this is partly what Popper was worrying. So the conclusion is, don't try to demarcate science from non-science, but try to demarcate scientists from non-scientists. I think this is a promising approach to this problem.

4. Question: It often is said that the hallmarks of science are explanation and prediction, and that these distinguish science from mere descriptive disciplines such as history. Evaluate the alleged centrality of both prediction and explanation to science. Illustrate both the explanation and the prediction portions of your discussion with examples from the biological sciences, earth sciences, astronomy, or astrophysics.


Are prediction and explanation both central to science? To answer the question, first we need to know what are explanation and prediction. According to Brush, philosophers and scientists use these words in quite different ways, but I take the question as philosophical one. Namely, explanation is to derive a fact already known to us from a theory (or to show that the fact is likely from the theory's point of view), and prediction is to derive a fact unknown to us from the theory. Of course this is a bit too crude way to put (especially as for explanation there are a variety of theories about the nature of explanation), but I think these crude definitions is enough for our present purposes.

As for explanation, there are abundant examples that a new theory provided a new organization or new explanation to known facts, and presumably was accepted because of it. Copernican theory provided a new organized view about planets. For example it naturally explained why Venus and Mercury do not go far away from the Sun. There are studies that the Copernican system did not provide a better prediction of the motion of planets at that time (e.g. Gingerich's study), so this natural explanation seems to be the reason some astronomists accepted the Copernican system. Watson and Crick's DNA model had strong influence on genetics mainly because the double helix seemed to explain (or, at least to suggest a way to solve the problems) how the genes can replicate themselves, and how the genes can transmit complex genetic information. These are good examples in which explanation played a significant role in theory acceptance.

But, on the other hand, it seems to me that there are fields in science in which explanation does not play that important role. For example, let us take taxonomy in biology. Linneous's taxonomic system neatly classified species into hierarchical system. The system was accepted by his contemporaries and is still used now. But it just classifies, and does not explain why species are classified in that way (Linneous himself thought that species in the same genus share some essence, but I do not see any explanatory power in this account). Obviously the acceptance of Linnean system was driven by other factors.

Another example I can come up with is Lyell's uniformitarian program in geology. Lyell claimed that when geologists explain geological changes, they can appeal only to the causes which actually work now. I think that the reason uniformitarianism was accepted by other geologists is not because it explains better (I think in many cases catastrophists could explain equally well), but because the program sounded more scientific than catastrophism. In both of these cases, acceptance of a new scientific theory was not driven by explanation.

As for prediction, again there are abundant examples that a theory seems to be accepted because of its successful prediction. For example, Boveri-Sutton's chromosome theory of heredity claimed that Mendelian alleles are placed on chromosomes. This theory made predictions like alleles do not always segregate randomly, but make groups the number of which is determined by the number of chromosomes of the species. This prediction was partly confirmed and created new anomalies which lead to Morgan's theory of `a thread of beads' view of gene. The prediction played an important role in the acceptance of the chromosome theory of heredity. Another example is the acceptance of Hess's sea-floor spreading hypothesis in 1960s, which prepared the acceptance of the plate tectonics theory. Hess's hypothesis said that the sea floors are spreading from mid-ocean ridges, and continents move because of the spreading. Hess's hypothesis together with paleomagnetic theory provided a novel prediction that the magnetization of sea floors should be symmetrical on either side of mid-ocean ridges (called Vine-Matthews-Morley hypothesis), and this novel prediction was confirmed by Pittman. This is supposed to be a decisive event to the acceptance of the plate tectonics theory. In both cases, the prediction played a central role.

But again, there are cases scientists choose theories by the considerations other than their predictive power. The acceptance of the Copernican system before Keplar's laws was a case of such acceptance. The Linnean taxonomy provides a case again, because it does not predict anything about the nature of species or about species we will find in the future.

Therefore, my conclusion is that even though both explanation and prediction play important role in science, they are not essential part.

6. Question: What is meant by the phrase "mechanization of the world picture"? Critically evaluate the claim that physics was mechanized in the seventeenth century, chemistry in the eighteenth, and biology in the nineteenth. Demonstrate detailed knowledge of at least one case study in each of the sciences in your answer.


The phrase `mechanization of the world picture' was used by Dijksterhuis to describe the change from Aristotelian world view to modern world view. According to him, a world picture is mechanized when people believe that the phenomena in the world can be and should be explained by mechanical laws, without appealing to the notions like `purpose', `design' and so on.

Physics is often claimed to be mechanized in 17th century. In Aristotelian world view, motions on the earth are divided into natural motions and unnatural motions, and each element are supposed to have a tendency to go back its own natural place. Through Galileo's and Descartes' contributions, Newton finally wiped out such teleological view about motion, and explained all motion by his three laws of motion and the law of universal gravitation. This is the ordinary understanding of the change, but close analysis of Newton's view shows that he had not overcome teleology entirely. Newton tried to explain his law of universal gravitation by some mechanical method, and failed. This is why he declared `I feign no hypothesis' and gave up explaining the law. But in the Leibniz-Clark correspondence, Clark (who is a student of Newton and who is supposed to represent Newton's view) claimed that the law of universal gravitation is not a mechanical law at all, because we need to assume constant intervention of God who makes sure that everything moves according to the law. Leibniz opposed to this interpretation saying that this makes the God a poor watchmaker, who cannot make a complete watch and who need to repair it from time to time. Now the plausibility of each claim is not our concern here. What we need to note is that both sides of the debate took God's design of the world for granted. Their mechanical world views were supported by the teleological presupposition. Therefore we can conclude that the mechanical world view was not totally attained in 17th century in physics.

In 18th century, a revolution took place in chemistry. Lavoisier effectively attacked phlogiston theory with his oxygen theory and established the law of conservation of matter. The nature of the revolution is often characterized as the replacement of somewhat mystical notion of `principles' with compound theory of matter. For many phlogiston theorists, phlogiston was a `principle' which has no weight (or, sometimes, negative weight, "levity") and add some property to matters to make it burn. The most revolutionary part of Lavoisier's theory was that oxygen , which make combustion possible, was an ordinary element with weight and volume. All chemical reactions are explained by combination and recombination of these ordinary elements. This is why Lavoisier's theory is regarded as mechanical. But closer examination of Lavoiser's theory shows that he had not overcome `principles' in his theory. According to his final version of the theory of combustion, oxygen gas was composed of oxygen and caloric, which is supposed to be an element of heat. In combustion of metal, oxygen is combined with the metal and release caloric. Lavoisier's caloric had the properties much like the principles in phlogiston theories. So we need reservation when we claim that Lavoiser mechanized chemistry.

In 19th Century biology, the major revolution was created by the publication of the Origin of Species by Charles Darwin, in 1859. Before Darwin, biologists thought that teleology is indispensable for the explanation of organisms. Even evolutionists like Erasmus Darwin or Lamarck had to assume that evolution occurs because of the design of God (Erasmus Darwin) or intentional use and disuse of characters by organisms (Lamarck). An influential anatomist, Cuvier thought that since the balance of organismic structure is so subtle, we need to assume the existence of Designer for the attainment of such a subtle balance. Darwin overturned this presupposition by introducing a totally mechanical way of explaining evolution of complex organisms. His theory of natural selection assumes two things: inheritable variables and struggle for existence among individual organisms. Because of the struggle, those organisms with advantageous variables are more likely to leave offspring than others. By the accumulation of this effect, the evolution takes place after a long time. In this explanation, we do not need any teleological view. This is a good case of mechanization.

However, when we claim that biology was mechanized in 19th century, we need to be careful. Darwin's theory has two parts, theory of transmutation and theory of natural selection. It is true that the theory of transmutation (the theory that all living organisms have common ancestors and have diverged by transmutation) was largely accepted after the publication the Origin, but the theory of natural selection was not met with that affirmative attitude. Many biologists thought that natural selection is not sufficient as the cause of evolutionary change. For the wide acceptance of the theory of natural selection, we need to wait until 1930s, synthesis of Mendelism and Darwinism by population geneticists.

My conclusion from these cases is that it is true that important theoretical events toward mechanization occurred in each case, but total appreciation of new view was not attained at that time. The appreciation of full implications of revolutionary views needs more time.